Malcolm Tredinnick wrote: > Cool. I'll have a look at it in detail. From a first read, though, it > looks like a reasonable approach: building in cache management to the > view. Why can't this be done using the normal CacheMiddleware, though? > I hadn't read the source on that; now I have. It's close, but it's not exactly the same.
I've spun myself into senselessness on this issue and I'm not sure I can defend it anymore. (Seriously; I've been running around with this for an hour now.) You can't quite get the *exact* behavior out of any combination of wrappers or middleware, because in general only the view can use local information to realize it can jump out early, as in the patch. Also in general, the view can make smarter decisions about ETags and Last-Modified than the middleware can. But if the cache middleware is enough to satisfy you, I'll just maintain my changes locally. (It is still the behavior I want; by the time you have "category feeds" and "all entry feeds" and a couple of other kind of feeds, trying to work out which feeds to expire when a blog entry changes is not worthwhile. This doesn't require fragile guessing of which keys to expire. I also actually *like* the fact this isn't dependent on the cache middleware at all; at the moment I'm not actually using memcached or anything because it doesn't make sense anywhere but RSS. It's the best solution for me. Ultimately, it's not that the built-in caching support can't do it at *all*, it's that a smarter view can do it *better*, but that may not be enough to justify breaking the rules. No sarcasm. I respect your collective judgments.) Broadly speaking, I'm still critical of a syndication framework that makes it easy to create RSS feeds that support no caching at all; many is the blogger that has been bitten by feeds that didn't support any of the caching options. Not every user can be expected to realize this. But perhaps the best solution is simply to document this potential problem clearly and wash your hands of it. > The work you seem to be doing may well be of real benefit, so don't > think I'm dismissing it out of hand. Even just talking through the > design is going to help somebody someday. > No, I understand. You can't just throw any old thing into the repository. I wouldn't like Django if you did. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django developers" group. To post to this group, send email to django-developers@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~----------~----~----~----~------~----~------~--~---
